Annada Behera

I research differentiable rendering, diffusion, and robust ML pipelines. I make large experiments reproducible and fast.

Currently, I am a PhD candidate and Senior Research Fellow at National Institute of Science Education and Research, Bhubaneswar under the supervision of Subhankar Mishra. My main area of competency is probability theory, diffusion, and automatic differentiation across large computational graphs like the rendering equation.

Contact me

The best way to reach me is on email, or Telegram. My office is M323, School of Mathematical Sciences, NISER, Bhubaneswar.

I generally hangout in #python, #voidlinux, #linux, #kakoune, #latex rooms of Libera Chat and in rooms #meta, #comfy, #tilde.green of tildeverse under the alias sticher or segfault. I read and sometimes engage in public threads on Usenet newsgroups, tilde.{green,meta,institute,services} and comp.{text.tex,lang.{c,awk,python},unix.{shell,misc}}. Come and say hi.

I have programming experience in

• managing headless Linux environment,
• writing models in PyTorch, doing training and inference on headless servers with SSH,
• environment management with uv including building against various libraries against various GCC versions and CUDA architecture,
• working on cloud computing infrastucture, particularly GCP,
• scene optimization with differentiable rendering (with Mitsuba and Dr.jit),
• novel view-synthesis with NeRFStudio,
• photogrammetry with Alicevision Meshroom API.
• training and inference of stable diffusion algorithms with Huggingface diffusers,
• image processing with OpenCV and Kornia,
• visualization with PyVista (VTK), Pyplot (Matplotlib), and

I have working knowledge of

• probability and information theory,
• approximations with Monte-carlo estimators and variance reduction using sampling techniques,
• automatic diffentiation across large computational graphs containing non-differentiable approximations like Monte Carlo estimators, with pushforwards and pullbacks.
• attention mechanism, encoders/decoders, embeddings, layers and their spectral analysis,
• working with text embeddings, VAEs and LoRA adaptes for stable diffusion, and
• various integrators for solving stochastic differential equations in diffusion.
• machine learing network architecture like AlexNet, UNet, VAE, and architecture design patterns,
• working with 3D rendering and shading pipelines,

Programming

• Programming languges: C/C++/CUDA (gcc, llvm, nvcc), Python, Odin
• Database: (SQL-database only) DuckDB, Postgres, Sqlite 3.
• Software applications: Blender, Meshroom, CloudCompare, Meshlab.
• Libraries: Numpy, PyTorch, Matplotlib, PyVista(VTK), OpenCV, Kornia.
• ML Libraries: Huggingface diffusers for stable diffusion, NeRFStudio (nerfacto, splatfacto) for 3D radiance field and Gaussian splat reconstruction, Mitsuba (differentiable ray tracer), Dr.jit (differentiable vectorization library),
• Cloud Compute: GCP
• Tech Stack: Linux (operating system), Git (version control), LaTeX (typesetting library), MetaPost (drawing library), PGF/TikZ (drawing library), uv (python environment), Fish (shell), Kakoune (text editor).

Talks and conference presentations

• August 2021, CUDA C Programming NISER, Bhubaneswar
• December 2021, Information theory and bayesian learning NISER, Bhubaneswar
• January 2022, Introduction to PyTorch with VAE NISER, Bhubaneswar
• June 2022, Irregular laten grid for shape reconstruction NISER, Bhubaneswar
• June 2022, Automatic diffentiation Summer Reading Group, NISER, Bhubaneswar
• July 2022, Differentiable rendering with edge sampling NISER, Bhubaneswar
• September 2022, Handling position and visibility discontinuities for physically-based differentiable rendering NISER, Bhubaneswar
• June 2023, Dr.Jit: A Just-In-Time Compiler for Differentiable Rendering PhD Pathana Gosthi 2023, NISER, Bhubaneswar
• September 2023, Implicit Neural Representations with Periodic Activation Functions NISER, Bhubaneswar
• November 2023, 3D Gaussian Splatting for Real-Time Radiance Field Rendering NISER, Bhubaneswar
• January 2024, Nitty-gritty details of the cuda rasterizer NISER, Bhubaneswar
• July 2024, Understanding the intuition and math behind Stable Diffusion NISER, Bhubaneswar

These are some talks I have presented in past, although some talks before 2021 are missing either because I didn't make or lost the files.

Peer-reviewed works

• Behera, Annada Prasad; Hait, Swati Rani; Dutta, Bapi; Mishra, Subhankar Restriction-based overlap and grouping functions with its application in Convolutional Neural Network architecture Information Sciences, 2025.
• Behera, Annada Prasad; Mishra, Subhankar Digitizing Temples for heritage conservation: Kalinga Architecture, Digital Applications in Archaeology and Cultural Heritage, 2025.
• Behera, Annada Prasad; Mishra, Subhankar Classical orthonormal polynomials as activation functions for implicit neural representations to preserve high frequency sharp features, 16th APCV, 2024
• Behera, Annada Prasad; Mishra, Subhankar Neural directional distance field object representation for uni-directional path-traced rendering, 14th ICCCNT, 2023. Published at IEEE Xplore.
• Joshi, Rucha; Behera, Annada Prasad; Mishra, Subhankar eBIM-GNN: Fast and scalable energy analysis, IEEE EnergyCon, 2022
• Joshi, Rucha; Behera, Annada Prasad; Mishra, Subhankar Temporal Motifs in Smart Grid, SMARTGREENS, 2020
• Routray, Ashanta R; Agarwalla, Santosh K; Behera, Annada Prasad Numerical Computing with Simulation of Basic Electrical Circuits, 2016 LAP LAMBERT Academic Publishing, ISBN-13:978-3659894398

Work experience and funded projects

• 2024-2025, ML Reseacher Engineer, PieLabs Inc.. My job was to benchmark various machine learning algorithms for automated testing of mobile apps from screenshots.
• 2023-2024, Senior Resarch Fellow, Tirtha: Crowdsourcing 3D hertitage models. Project under Project under Subhankar Mishra, funding from La Fondation Dassault Systemes, NISER, Bhubaneswar. My job was to create Gaussian splats of heritage sites.
• 2021-2022, Senior Resarch Fellow, Digitization of Rajarani and Lingaraj temples. Project under Project under Subhankar Mishra, funding from La Fondation Dassault Systemes, NISER, Bhubaneswar. My job was to develop methods to efficiently store, transmit and exhibit huge photogrammetry and 3D point cloud data.
• 2019-2021, Junior Resarch Fellow, AutoBIM: BIM to GIS Integration. Project under Subhankar Mishra, funding from DST-NRDMS, Government of India, NISER, Bhubaneswar. My job was to create extract 3D geometry of buildings for the integration into BIMs.
• 2018, Intern, Theoretical and Empirical Analysis of Machine Learning Algorithms: Meta-heurristic Optimization, Internship under guide Prof. Ashish Ghosh, Center for Soft Computing Research, Indian Statistical Institute, Kolkata I worked to test various meta-heurristic optimization strategy.
• 2016, Embedded Systems and Internet of Things Intern in 2016, Hetrogenous Communication Technology Pvt. Ltd, Guindy, Chennai. I worked in testing software applications controlling IoT systems.

National level qualifications

• Graduate Aptitude Test in Engineering (GATE), 2020 Computer Science and Information Technology (CS&IT)
• UGC National Eligibility Test (UGC-NET), 2019 Computer Science and Applications (CSA)
• Graduate Aptitude Test in Engineering (GATE), 2019 Computer Science and Information Technology (CS&IT)
• Joint Entrance Screening Test (JEST), 2019 Theoretical Computer Science (TCS)
• Odissa Joint Entrance Examination (OJEE), 2017, topped Master in Computer Application (MCA)

Conference and workshops participated

• 1st DAE Conclave, 2024 The Annual Convention of Department of Atomic Energy Scientists and Engineers, October, National Institute of Science Education and Reseach, Bhubaneswar, India.
• APCV 2024 SG The 16th Asia Pacific Conference on Vision, July, Yale-National University of Singapore, Singapore.
• ARCS 2024 Academic Reseach and Careers for Students, Association for Computing Machinary (ACM), February, NISER, Bhubaneswar, India.
• AI&TCS'23 Winter School on Primer on 3D Vision In AI 2023, December, Indian Institute Technology, Jodhpur, Rajasthan, India.
• InHERIT 2023 International Heritage and Cultural Conservation Conference August, University of Technology, Sarawak, Malayasia.
• ICCCNT 2023 14th International Conference on Computing, Communication and Networking technologies July, Indian Institute of Technology Delhi, India.
• ACM CODS-COMAD 2023 6th Joint International Conference on Data Science & Management of Data January, Indian Institute of Technology Bombay, Mumbai, India.
• NVIDIA GTC Nvidia GPU Technology Conference, June 2022, Online.
• Microsoft Resarch India Academic Summit, January 2020, BITS-PILANI, Goa, India

Teaching

• Computational Laboratory-I, 2022 teaching assistant for Anup Bhattacharya at NISER, Bhubaneswar.
• Programming and Data Structures Lab-II (CS142) teaching assistant for Anup Bhattacharya at NISER, Bhubaneswar.
• CS141 - Programming in Python, 2021, teaching assistant for Subhankar Mishra NISER, Bhubaneswar.
• CS 142 - Programming and Datastructures II, 2021, teaching assistant for Subhankar Mishra NISER, Bhubaneswar.
• Computational Laboratory-I, 2020, teaching assistant for Rishiraj Bhattacharyya at NISER, Bhubaneswar.
• Computational Physics, instructor under Odissa Higher Education Program for Excellence and Equity with funding from World Bank at Fakir Mohan University, Balasore in 2019.

Last updated: Mon Sep 15 05:54:43 PM IST 2025